The Effect of Training, Compensation and Work Discipline on Job Satisfaction that Impacts Employee Performance PT. Astra Honda Motor

Received: 01 October 2021 Accepted: 22 October 2021 Published: 24 November 2021 DOI: 10.32996/jbms.2021.3.2.23 This study aims to determine the effect of training, compensation, and work discipline on job satisfaction that has an impact on the performance of employees of PT. Astra Honda Motor. This research uses a case study method with a quantitative research methodology that uses 200 respondents from PT. Astra Honda Motor in the Engineering Plant Division 1 to Plant 5, Purchasing Division (P & PC) Plant 2, Procurement Plant 2 Division, PPIC Plant 1 Division to Plant 5. In this study, the Likert scale was used to model the questionnaire. The statistical testing technique in this study is multiple linear regression and hypothesis testing using the SPSS series 22 program. Meanwhile for mahalanobise distance test, path analysis, and trimming use the Amos series 22 program. KEYWORDS


Introduction 1
PT. Astra Honda Motor (AHM) is a pioneer in the motorcycle industry in Indonesia, which was established on June 11, 1971, with the initial name PT. Federal Motors. Currently PT. Astra Honda Motor has 20,000 employees. To face increasingly competitive business competition, companies need to manage human resources well. So we need employees who have competence and integrity in their work. Several factors are required to improve employee performance, including training, compensation, and work discipline factors associated with job satisfaction factors. The purpose of this study was to determine the significance of the effect of training, compensation, and work discipline on employee performance mediated by job satisfaction on employees of PT. Astra Honda Motor.

Relationship Between Variables and Research Hypotheses 1.Effect of Training on Job Satisfaction.
The effect of training on job satisfaction is found in the results of the training programs provided by the company to employees. Because with effective and efficient training programs that are held on an ongoing basis, it will be able to improve employee competencies. So that employees can perform activities well, and this can lead to job satisfaction. Then this is reinforced by several previous research results, including:

Effect of Compensation on Job Satisfaction.
The effect of compensation on job satisfaction will provide enthusiasm and work motivation significantly. Because compensation is the main factor in terms of employee satisfaction, so that employees will feel valued by the company for the contribution of energy, thought, and expertise provided by employees to the company. This is further strengthened by several research results, including: 1. Research conducted by Gopinath (2016) entitled "How The Compensation Management and Welfare Measure Influence Job Satisfaction? A Study with Special Reference in Bsnl to Three Different Ssas Using Modeling". With research results, BSNL has a good compensation management system, which is equivalent to industry standards. School Teachers in Maara Sub -County of Tharaka Nithi County, Kenya". The results of this study found that compensation is one of the incentives that motivate workers.
Then the research hypothesis on the effect of compensation on job satisfaction, namely:

H2: There is a positive and significant effect between compensation (X2) on job satisfaction (Z) on employees of PT. Astra
Honda Motor.

Effect of Work Discipline on Job Satisfaction.
The effect of work discipline (X3) on job satisfaction (Z) positively impacts the company. Because employees who have a high level of discipline will always do work with pleasure so that it will directly affect job satisfaction. Then this is reinforced by several previous research results, including:

Effect of Compensation on Employee Performance.
The effect of compensation on employee performance will provide significant work motivation. Because compensation, according to researchers is a company strategy to be able to significantly improve individual performance and be able to retain employees, so that it will reduce the intensity level of employee turnover. Then this is reinforced by several previous research results, including:

Effect of Work Discipline on Employee Performance.
The effect of work discipline on employee performance will result in good productivity. Because in general, employees who have a high level of work discipline will work using quality and quantity following the company's SOPs. Furthermore, this is reinforced by several previous research results on the effect of work discipline on employee performance, as follows: Then the research hypothesis on the effect of work discipline on employee performance, namely:

H6: There is a positive and significant influence between work discipline (X3) on employee performance (Y) at PT. Astra
Honda Motor.

Effect of Job Satisfaction on Employee Performance
The effect of job satisfaction on employee performance will result in good individual performance. Because if employees feel satisfied while working, it will undoubtedly cause feelings of pleasure in carrying out their work activities. Then this is reinforced by several previous research results, including: 1. Hittiarachi (2014) research entitled "Impact of Job Satisfaction on Job Performance of IT Professionals: With Special Reference to Sri Lanka." With the research results, the job satisfaction dimensions (salary, promotion, supervision, and work) on software engineers positively affect (16.7%) so that engineers are satisfied with their work. 2. Research conducted by Inuwa (2016) entitled "Job Satisfaction and Employee Performance: An Empirical Approach". With the results of research job, satisfaction has a positive and significant influence on employee performance. This clearly indicates an increase in the level of job satisfaction of non-academic BASUG staff will also lead to improvement and higher performance.
Then the research hypothesis on the effect of job satisfaction on employee performance, namely:

H7: There is a positive and significant effect between job satisfaction (Z) on employee performance (Y) on employees of PT. Astra Honda Motor.
The conceptual model in this study aims to design research systematically. The conceptual model is a diagram of causality between certain factors that are believed to have an impact on an object of research. The framework of thinking aims to provide a systematic description of the object of research. The data collection technique in this study aims to obtain the data needed by researchers to analyze the phenomena or problems in this study. The primary data in this study came from a questionnaire using a Likert scale technique. There are three types of variables in this study, namely the independent variable (exogenous), the mediating variable (intervening), and the dependent variable (endogenous).
1. Exogenous Variable aims as a variable that affects the endogenous variable (the dependent variable), which is the factor that is measured by the researcher to determine the relationship between the observed phenomena. a) X1 = Training, b) X2 = Compensation, c) X3 = Work discipline. 2. Intervening Variable aims as a variable that affects the correlation between exogenous variables (independent variable) and endogenous variables (dependent variable). Z = Job satisfaction. 3. Endogenous Variable aims as the dependent variable, namely the observation factor, and is measured to determine the relationship of exogenous variables, namely the factors that arise. Y = Employee Performance.
Operational variables in this study are used to describe research variables, dimensions, and indicators.

Pre-test Analysis 2.1.1 Validity Test
The test of the research data instrument in this study is to compare the value of r-count with r-table for the degree of freedom (df) = n-2, the sample in this study amounted to 50 respondents, then a correlation analysis was carried out between the statement score and the total score with the criteria if the statistical probability value > 5% significance level = 0.05 then the item is declared valid.

Reliability Test
Reliability is a measuring tool that shows the stability of the data results and the possibility that the same researcher uses the measuring instrument at different times and vice versa, or at other times by using the basis for decision making using the Cronbach alpha technique because the research instrument has the form of a questionnaire with a Likert scale.
With an explanation, if alpha > 0.90, it indicates perfect reliability. If the alpha value is between 0.60 -0.90, it means high reliability. If the alpha value is 0.50 -0.70, it means moderate reliability. If the alpha value < 0.50 means low reliability. If the alpha value is low, it is possible that one or more items are not reliable.

Main Test Analysis
Respondents used in this research amounted to 400 people. In this study, the data testing technique used the following methods: a. Multiple linear regression with classical assumption tests including normality, multicollinearity, heteroscedasticity, and linearity. b. The test of mahalanoise distance. c. Path analysis (path analyst). d. Trimming test. e. Hypothesis test (t-test and R²)

Validity Test
A validity test is a measure that shows the level of validity or validity of an instrument using 50 respondents. The results of this study are as follows:

Reliability Test
A construct or variable is said to be reliable if it gives a Cronbach alpha value > 0.60 as follows: Based on the results of the data table 3.2, the researcher concludes that all variables in this study have a significance level of >5% or 0.05, which means that all of these variables are declared reliable and feasible to be tested for the primary test.

Classical Assumption Test.
The most appropriate regression parameter is obtained using the OLS (Ordinary Least Square) method of least squares to get the equation. The OLS method can be used as an unbiased estimate if it meets the BLUE (Best Linear Unbiased Estimation) requirements. Based on the data in table 3.3, it is known that the c.r. (critical) skewness or c.r (essential) kurtosis is between -2.58 and 2.58. So it can be concluded that all data in this study are normally distributed, including the results of the multivariate numbers, because all data are below 2.58 and above -2.58. Based on the results of the data table 3.4, it is known that each exogenous variable has collinearity statistics > 0.1 and VIF tolerance < 10. Then it can be concluded that all of the above variables do not occur multicollinearity.  Figures 3.1 and 3.2, it can be concluded that the points spread with an unclear pattern and are below the number 0 on the Y-axis. Therefore, the researcher concludes that there is no heteroscedasticity problem in the substructure 1 and substructure 2 regression models. Table 3.

The Effect of Training, Compensation and Work Discipline on Job Satisfaction that Impacts Employee Performance PT. Astra Honda Motor
Page | 226 ANOVA Based on the data, table 3.5 shows that all variables have a significant value of deviation from linearity <0.05. It can be concluded that all variables have a linear relationship so that the linearity assumption is met. Based on the results of the data table 3.6, no outlier data was found, because the values of p1 and p2> 0.05. Therefore, the mahalanoise distance test in this study was fulfilled.

Path Analyst
The direct and indirect effects on path analysis in this study are as follows: Based on the data from table 3.7, the researcher states the category of the influence of the relationship between the other variables (p) being studied where if the path coefficient is 0.05-0.09, then the influence power is included in the weak category, if the path coefficient is 0.10-0.29 then the power is categorized as weak. the influence is in the medium category; if the path coefficient is > 0.30 then the impact is in the medium category. Therefore, the researcher describes as follows: 1. The variable directly influences job satisfaction with an effect of 0.194 or (19.4%) and a correlation coefficient of 0.223. This value shows the magnitude of the influence is low and accurate. 2. The compensation variable directly relates to job satisfaction with an influence magnitude of 0.250 or (25.0%) and a correlation coefficient value of 0.259. This value shows the extent of the influence is low and real. 3. The work discipline variable directly correlates with job satisfaction with a magnitude of 0.141 or (14.1%) and a correlation coefficient of 0.223. This value shows the magnitude of the influence is very low and real. 4. The variable has a decent direct relationship of 0.178 and a decent indirect relationship of 0.043 on employee performance with a total effect of 0.221 or (22.1%) and a correlation coefficient value of 0.207. This value shows the magnitude of the influence is low and real. 5. The compensation variable has a direct relationship worth 0.278 and an indirect relationship worth 0.055 on employee performance with a total effect of 0.333 or (33.3%) and a correlation coefficient value of 0.334. This value shows the magnitude of the influence is low and real. 6. The work discipline variable has a decent direct relationship of 0.100 and a decent indirect relationship of 0.031 on employee performance with a total effect of 0.131 or (13.1%) and a correlation coefficient value of 0.165. This value shows the magnitude of the influence is low and real. 7. The job satisfaction variable has a direct relationship worth 0.221 to employee performance with a total effect of 0.221 (22.1%) and the correlation coefficient value is 0.301. This value shows the magnitude of the influence is low and real.
Then seen from the results of the data table 3.7, it can be described using the IBM AMOS series 22 program as follows:  Based on the significance test results for the job satisfaction variable, the t-count value was 4.502, with a significance level of 0.00 <0.05. Then the researcher concluded that H7 was accepted and H0 was rejected. Based on the data in table 3.9, it can be seen from the acquisition of the R-Square value of 0.120 or (12%). This is means that exogenous variables (training, compensation, and work discipline) simultaneously affect job satisfaction by 0.120 or (12%). In comparison, the rest (88%) is influenced by variables not discussed in this study, such as organizational commitment, development, motivation, and employee retention. Based on the data in table 3.10, it can be seen from the acquisition of the R-Square value of 0.209 or (20.9%). This means that exogenous variables (training, compensation, work discipline, and job satisfaction) simultaneously affect employee performance by 0.209 or (20.9%). While the rest (79.1%) is influenced by other variables not discussed in this study, such as organizational commitment, development, motivation, and organizational culture.

Conclusions
This study uses a case study that functions as a problem-solving test by analyzing the effect of training, compensation, work discipline on employee performance mediated by job satisfaction of employees of PT. Astra Honda Motor. The data collection technique in this study used a questionnaire with a Likert scale calculation and used descriptive interview techniques, namely interview techniques by approaching. Respondents in this study were employees of PT. Astra Honda Motor totaled 400 respondents. Based on the results of the study, the following results were obtained: 1. The effect of training on job satisfaction. From the study results, it can be concluded that there is an influence between activity and job satisfaction. 2. The effect of compensation on job satisfaction. From the study results, it can be concluded that there is an influence between compensation and job satisfaction. 3. The effect of work discipline on job satisfaction. From the study results, it can be concluded that there is an influence between work discipline and job satisfaction.
4. The effect of training on employee performance. From the study results, it can be concluded that there is an influence between training on employee performance. 5. The effect of compensation on employee performance. From the results of the study, it can be concluded that there is an influence between compensation on employee performance. 6. The effect of work discipline on employee performance. From the study results, it can be concluded that there is an influence between work discipline on employee performance. 7. The effect of work discipline on employee performance. From the results of the study, it can be concluded that there is an influence between work discipline on employee performance.